22MAY

YouGov second Welsh MRP flips Plaid lead

3 min read

10:09UTC

YouGov's second 2026 Senedd MRP, on fieldwork from 6-15 April, projects Reform UK on 37 seats and Plaid Cymru on 36, reversing a 13-seat Plaid lead from six weeks ago, with both parties on identical 29% vote share.

← Back to Three constitutional contests open Jump to analysis ↓

PoliticsDeveloping

Key takeaway

The Welsh majority arithmetic the left was counting on six weeks ago no longer exists.

YouGov's second 2026 Senedd MRP, with fieldwork from 6 to 15 April on 2,387 Welsh adults, projects Reform UK on 37 seats and Plaid Cymru on 36 ¹. Both parties share 29% of the vote on central estimates. Welsh Labour is unchanged on 12, the Wales Greens drop from 10 to 7, the Welsh Conservatives rise from 1 to 3, and the Welsh Liberal Democrats take 1.

The first MRP, in March, had Plaid on 43 and Reform on 30 . The seat swing across six weeks is fourteen: seven from Plaid to Reform, plus three lost from The Greens. Yet the same fieldwork puts both parties on identical 29% vote share, suggesting the swing is geographic rather than aggregate. The Plaid-Green coalition arithmetic that opened the left's route to a 49-seat majority is no longer in the model. A Plaid-Green combination now reaches 43 seats, six short. A Plaid-Labour combination reaches 48, one short. All three left-leaning parties together carry a majority in 96% of simulations, but require active cooperation among parties already arguing publicly about who counts as left in Wales .

Two other models broadly corroborate the closeness without matching the YouGov reversal. Ipsos's Wales poll on fieldwork to 8 April puts Plaid on 30%, Reform 25, Labour 15. More in Common's Senedd MRP projects Plaid 30, Reform 28, Labour 24, Greens 4 ². Three independent models agree on one thing: no party wins a Senedd majority alone. They disagree on which of Reform or Plaid finishes first, and on whether The Greens cross the entry threshold.

Wales is using closed-list proportional representation for the first time , with 16 six-member constituencies and the D'Hondt allocation method (votes divided by 1, 2, 3 etc., seats awarded to highest quotients). Closed-list PR registers small movements between adjacent parties contesting the same ideological space sharply. The 7-seat shift from Plaid to Reform reflects under three points of Welsh vote share moving across that line. The system is doing what it was designed to do; Welsh voters are watching it for the first time.

Deep Analysis

In plain English

Welsh voters on 7 May will use a new voting system for the first time. Instead of voting for a local candidate, they now choose a party list across 16 large areas, and seats are divided up in proportion to how many votes each party gets in each area. YouGov, a polling company, ran an MRP model , a technique that combines a large survey with local demographic data to project seat counts. The first model in March projected Plaid Cymru (the Welsh nationalist party) winning 43 seats and Reform UK winning 30. The second model, from mid-April, reversed this: Reform 37, Plaid 36, both on 29% of the vote. To form a Welsh Government, a party or group of parties needs 49 seats. Under the first model, Plaid plus the Wales Greens reached 53, comfortably above that threshold. Under the second model, Plaid plus the Greens reach only 43, six seats short. A Plaid government would then need Welsh Labour to join as a partner. The 33-vote gap between models is within the statistical margin of error for the survey size, meaning both outcomes are genuinely possible.

Deep Analysis

Root Causes

Wales's new closed-list proportional representation system, introduced for the 2026 election under the Senedd Cymru (Members and Elections) Act 2024, has no prior electoral data for MRP calibration. D'Hondt allocation across 16 six-member constituencies means seat totals are highly sensitive to small vote-share movements in specific regions, amplifying the variance between model runs.

The Plaid-Reform closeness reflects the Wales Governance Centre's February 2026 finding that Welsh political realignment is a bloc consolidation rather than a cross-party swing: Labour-to-Plaid movement in the Welsh/progressive bloc and Conservative-to-Reform movement in the British/right bloc are occurring simultaneously. Both blocs reached near-identical size in the second MRP, meaning the election outcome depends on which bloc's voters turn out at higher rates.

What could happen next?

Consequence
If the second MRP is accurate, Plaid Cymru cannot form a government without Welsh Labour, transforming the post-election negotiation from a binary Plaid-Green agreement to a three-party or two-party Plaid-Labour coalition.
Risk
A Reform plurality at 37 seats , without a majority path , produces a hung Senedd where every potential government requires cooperation across the Welsh/British bloc divide, a scenario the Wales Governance Centre has described as constitutionally untested.

Source Landscape

This story draws on neutral-leaning sources

Primary parallel: The 2015 UK general election saw YouGov's final MRP project a hung Parliament; the Conservatives won an outright majority. Post-election analysis by YouGov's own team attributed the error to differential turnout models in English marginals , the same structural problem that Welsh closed-list constituencies present, where party-list vote is more uniform across socioeconomic groups than constituency first-past-the-post polling.

The 2026 Welsh MRP faces the same methodological challenge: turnout in first-ever PR contests in communities with historically high Labour turnout is genuinely unknown.

Counter-parallel: The 2021 Senedd election was accurately called by YouGov's final MRP within two seats of the actual result. That model had been calibrated against a previous PR election (the national assembly had run AMS since 1999), while 2026 uses a new pure closed-list system (ID:2109) with no historical equivalent for model calibration.

Consensus view: The Wales Governance Centre at Cardiff University, led by Roger Scully, has tracked Welsh MRP volatility across the 2026 cycle and notes that a 13-seat swing between two YouGov models three weeks apart reflects genuine polling uncertainty rather than a real shift in voter intention: the margin of error on a 2,387-person sample across 16 closed-list constituencies is wide enough that both models are statistically consistent with a razor-thin Reform or Plaid lead.

Scully's team has described the election as the most volatile in Welsh polling history.

Counter-view: The Bevan Foundation's Steffan Evans argues that the MRP reversal captures a real late-campaign shift in south-east Wales coastal constituencies, where Reform's canvassing has been concentrated, and that treating both models as statistically equivalent understates genuine momentum evidence from ground-level data. His analysis identifies Rhondda Cynon Taf and Blaenau Gwent as the specific seats where the swing is structurally plausible.

Key tension: Whether the Reform lead in the second MRP reflects genuine movement or methodological noise, which will not be resolved until counts begin on 8 May.

Sources:YouGov·More in Common

Mentions:YouGov →Reform UK →Plaid Cymru →Wales Greens →Welsh Labour →More in Common →Ipsos →Welsh Conservatives →Prima →Wales →Senedd →Wales Governance Centre →Liberal Democrats →Cardiff University →Green Party →

First Reported In

Update #5 · 11 Days to Go: Six-of-six, RPA dies, Welsh lead flips

YouGov· 26 Apr 2026

Read original →

Causes and effects

Caused by

Wales Greens projected as Senedd kingmakers

The first YouGov Senedd MRP (2316) showing Plaid 43, Reform 30 was reversed by the second MRP projecting Reform 37, Plaid 36.

Occurred 11 Apr 2026

Read story →

Plaid and Wales Greens trade public blows

The Plaid-Wales Greens public dispute (2419) is structurally consequential given the second MRP closes the Plaid-Green majority route.

Occurred 14 Apr 2026

Read story →

This Event

YouGov second Welsh MRP flips Plaid lead

Wales is using closed-list proportional representation for the first time, where small national vote-share movements convert directly into seat allocation; the seat split between Reform and Plaid is now genuinely unknown.

Led to

Wales Greens fall from 10 to 2

YouGov's second Senedd MRP (Reform 37, Plaid 36) was reversed by the final MRP (Plaid 43, Reform 34), flipping the projected coalition arithmetic

Occurred 5 May 2026

Read story →

Different Perspectives

UK Government (Labour)

Westminster framed the youth justice transfer as a culmination of prior work rather than a precedent, refused a Section 30 order before the request arrived, and omitted both the Representation of the People Bill and any Wales Bill from the 13 May King's Speech. Starmer is described as open to a devolved-leaders summit in June.

Reform UK

Essex, Norfolk and Suffolk filed pre-action protocol letters framing their LGR challenge as mandate-consistent, while 22 Reform councillors departed in 14 days at an annualised rate nearly three times the 10 percent projection. Richard Tice defended the Harborne 5 million pound gift as unconditional, with no acknowledgement of Farage's two contradictory accounts on record.

SNP (Scottish Government)

Swinney submitted a Section 30 request on 14 May citing the 73-seat SNP-Greens pro-independence bloc and the 2014 Edinburgh Agreement precedent, despite the SNP finishing seven below his self-set 65-seat trigger. Downing Street's contradictory readout of their call suggests Westminster is treating the request as a holding item rather than a live constitutional negotiation.

Welsh Government (Plaid Cymru)

ap Iorwerth framed the youth justice transfer as a starting point for his six-power Wales Bill agenda, pressing demands at a phone call with Keir Starmer on 18 May and winning the first statutory function transfer to Cardiff since 1999. The Greens' unwritten confidence-and-supply arrangement gives him 45 of 96 seats, four short of a majority.

Russell Findlay (Scottish Conservatives)

Findlay refused to resign as Scottish Conservative leader after the party fell to 12 Holyrood seats and lost all five constituency MSPs. He declined Swinney's post-election talks invitation, the only major-party leader to do so.

John Swinney (SNP)

Swinney committed on 14 May to a Holyrood Section 30 vote within a week despite winning seven seats fewer than his own trigger threshold, relying on a SNP-Green majority of 73. He tabled a meeting with Starmer; Downing Street disputed that any referendum discussion was agreed.